1590
BIRCH, DEAN, HUNTER, AND WHITEHEAD
mine-free product, 3,4-dinitrothiophene (111), if it were formed we prepared a sample of the material by the reaction between the bromodinitrothiophene I1 and copper in boiling butyric acid.8 3,4-Dinitrothiophene, which has already been prepared in similar fashion from the dinitrodibromothiophene I by Dr. Ellis Brown and which will be described in detail by him later, melts at 94-95' and mixtures of the monobromodinitrothiophene I1 and the dinitrothiophene I11 melt below 75". Z-~romo-S,$,~-~rinitr~thiophene (IV). Twenty grams (0.08 mole) of the crude bromodinitrothiophene I1 was added with stirring during 5 min. to a mixture of 100 ml. of concentrated sulfuric acid and 100 ml. of fuming nitric acid (d. 1.59-1.60). The solid dissolved in the nitrating acid, which was at room temperature, without an appreciable evolution of heat. I n about 0.5 hr. the trinitro compound began t o precipitate. Stirring was continued for 2 hr., after which time the reaction mixture was left overnight before i t was poured onto 300 g. of ice. The yield of crude nitration product, melting at 123-125", was 21.5 g. (90%). 2-Bromo-3,4,5-trinitrothiopheneis a pale yellow solid which is soluble in the common solvents, but apparently with solvolysis; the solutions develop much color and the solute crystallizes poorly or not a t all. The only satisfactory procedure for purification is to dissolve the trinitrobromothiophene in concentrated nitric acid (10 ml./g.) by heating on the steam bath. On cooling the filtered solution, the product is obtained with an 80% recovery as lemon yellow crystals melting a t 130-131". Anal. Calcd. for C4BrN80BS:C, 16.10; H, 0.0; Br, 26.8. Found: C, 16.16; H, 0.0; Br, 26.7. The trinitrobromothiophene furnishes only intractable material on treatment with potassium acetate in acetic acid. With aniline in methanol i t yields a dianilino derivative (V),
[CONTRIBUTION FROM
THE
VOL.
22
which is obtained as orange needles melting a t 196-196.5' after crystallization from acetone-methanol. Anal. Calcd. for C I ~ H ~ B N ~C, O ~53.9; S : H, 3.39. Found: C, 54.05; H, 3.46. Z,S,4-Trinilrothiophene (VI). A solution of 5.0 g. (0.017 mole) of the crude bromotrinitrothiophene IV in 100 ml. of acetone was cooled to about IO" and 10 ml. (0.85 mole) of 50% hypophosphorous acid was added. An orange-red color developed at once. The solution was kept at room temperature by immersing the flask in cold water from time to time. After 1 hr. the solution was poured into 800 ml. of water and the tan precipitate was filtered; y.ield, 2.8 g. (760/,), m.p. 138-141". The product was purified by crystallization from ether (30 ml./g.) and an equal volume of' 30-60' petroleum ether. The recovery was 70%. Anal. Calcd. for CdHNd&S: C, 21.90; H, 0.46. Found: C, 21.93; H, 0.60. 2,3,4-Trinitrothiophene is a colorless solid which melts a t 143'. When a small sample of the material in a sealed melting point tube is heated in a free flame i t explodes vigorously. The material, which analpzes satisfactorily for carbon and hydrogen, gives low values for nitrogen and sulfur. The trinitrothiophene is unaffected by the ordinary nitric acid-sulfuric acid nitrating mixtures a t steam bath temperatures. When one gram of the material was heated a t 130" for 1.5 hr. with 10 ml. of concentrated sulfuric acid and 2 g. of potassium nitrate or for 4 hr. with 5 ml. of nitrating acid of the Composition 82.3% H#04-17.6% HNOaO.l% HzO, there was no indication that nitration had taken place and 60% of esscntially pure starting material was recovered.
FLUSHING 67, N. Y.
RESEARCH STATION,THEBRITISHPETROLEUM COMPANY LIMITED]
Preparation and Physical Properties of Sulfur Compounds Related to Petroleum. VII. 2-, 6- and 8-Thiabicyclo[3.2.1]octane and 2-Thiabicyclo[2.22Ioctane STANLEY F. BIRCH, RONALD A. DEAN, NEVILLE J. HUNTER,
AND
EDhlUND V. WHITEHEAD
Received April 8,1957
2-, 6-, and 8-Thiabicyclo [3.2.l]octane and 2-thiabicyclo [2.2.2]octane have been synthesized and their physical properties have been recorded. Each sulfide has been characterized by the preparation of derivatives.
I n an earlier paper', Birch et al. reported the isolation from an Agha Jari (S. Persia) kerosine of a sulfide which was apparently 8-thiabicyclo[3.2.1]octane and suggested that other thiabicyclooctanes were also present in this kerosine. It has previously been found possible to identify certain other sulfides as constituents of the kerosine, by comparison of the infrared spectra of the isolated sulfides or sulfide mixtures with those of synthetic compounds which have been prepared in these laboratories.2 Consequently, the synthesis of 2-, 6- and 8-thiabicyclo[3.2.l]octane (IX, XI11 and IV) and 2(1) S. F. Birch, T. V. Cullum, R. A. Dean, and R. L. Denyer, Ind. Eng. Chem., 47, 240 (1955).
(2) S. F. Birch, R. A. Dean, N. J. Hunter, and E. V. Whitehead, J. Org. Chem., 2 0 , 1178 (1955).
thiabicyclo[2.2.2]octane (XVII) was carried out, as reported in the present paper. 3-Thiabicyclo[3.2.l]octane, the fifth member of the group of thiabicyclooctanes containing only five- and sixmembered rings and having an atomic bridge, has already been described. The stereochemistry of the thiabicyclo[3.2.1]- and thiabicyclo[2.2.2]0~tane systems is such that the trans- forms would be extremely strained and, as with the corresponding hydrocarbons, it is probable that only one form having a cis- configuration can exist.4 As recently (3) S. F. Birch and R. A. Dean, Chem. Ber., 585, 234 11954) in which this sulfide is referred to as 6-thiabicvcloii.2.3joctrtne.
(4) H. Gilman, Organic Chemistry, 2nd. Ed., John Wiley and Sons, Inc., New York, 1948, Vol. 1, p. 485.
DECEMBER
1957
SULFUR COMPOUNDS RELATED TO PETROLEUM. VI1
reported,6 it has been confirmed that all five of the above thia,bicyclooctanes are in fact present in the kerosine under investigation. These “atomic-bridge” bicyclic sulfides have not been synthesized previously, although the corresponding hydrocarbons and certain nitrogen and oxygen analogues have been prepared. In common with other bicyclic sulfides which we have synthesized2J they were obtained by two general methods: (a) hydrogenation and reduction of the unsaturated sulfones obtained by condensation of sulfur dioxide and the appropriate diene and (b) cyclization of the appropriate dibromide or ditosylate’ with sodium sulfide. The yields obtained by method (b) were rather poor but method (a), which gives better yields, can be used only for compounds in which the sulfur atom is part of a five-membered ring, and then only when the structure of the intermediate unsaturated sulfone would not violate Bredt’s Rule.8 Accordingly, of the required sulfides, only 8-thiabicyclo[3.2. lloctane (IV) was prepared by method (a); the reaction route is shown schematically (I-IV).
cc-c Q 2-,!s I (4I c=c-t: 6-b-t: c=c-c
1
4
cc-c (IV)
$
p
(11)
Hs(Ni)
C C C LiAIH4
+ ,!s
I
A I
(111)
1591
tosylate group of the intermediates used in this method, is accompanied by Walden inversion.9?l0 Accordingly, on reaction with sodium sulfide, a cis-intermediate should give a trans-sulfide, and a trans-intermediate a cis-sulfide. Since the formation of a trans-sulfide is prohibited the cis-intermediates actually give entirely polymeric material. Moreover, even with the trans-isomers, the formation of the required sulfides is hindered by the tendency for the secondary bromide or tosylate groups to eliminate with a neighboring ring hydrogen atom11112and, possibly, by the difference in the reactivities of the primary and secondary bromide or tosylate groupsgJOwhich may favor intermolecular reaction. It is thus not surprising that poor yields were obtained in these preparations. The dibromides and ditosylates required as intermediates in method (b) are, in general, most conveniently obtained from the corresponding diols by hydrobromination and treatment with p-toluenesulfonyl chloride respectively. Rearrangements usually occur during hydrobromination of secondary a l c o h o l ~resulting ~ ~ ~ ~ ~in the formation of mixtures of isomeric dibromides. Consequently, in the dibromide route tedious separation processes may be required a t either the dibromide or sulfide stage and/or there may be some uncertainty as to the structure of the pure sulfide finally isolated. On the other hand, tosylation of secondary alcohols takes place without isomerization16 and with re, ~ ~ ~the ~ structure of a tention of c o n f i g u r a t i ~ n and sulfide obtained via the ditosylate is thus unambiguous. However, this advantage may be offset if the parent diol of the required ditosylate is not easily prepared while the corresponding dibromide can be obtained from another compound. In addition, better yields are obtained by the dibromide
1,3-Cycloheptadiene (I) reacted with sulfur dioxide at a lower temperature (20”) and gave a better yield (95%) of the corresponding unsaturated sulfone than did 1-vinylcyclopentene in the analogous preparation of cis-2-thiabicyclo[3.3.0]octane.2 Reduction of the unsaturated sulfone (11) proceeded smoothly and no poisoning of the catalyst was experienced during this reaction. I n VI V common with other saturated bycyclic sulfones2; 8-thiabicyclo[3.2.l]octane-8,8-dioxide(111) was incompletely reduced by lithium aluminum hydride (2 moles) and gave only 65% of the expected sulfide together with 30y0 of unreduced sulfone. No + c-c-c isomerization2 occurred during this reduction, since the degree of strain required in the transisomer of the sulfone precludes its formation. The sulfide was purified by crystallization and sublimation. (9) M. F. Clarke and L. N. Owen, J . Chem. Soc., 2108 The other three sulfides were obtained by method 11950). (b). The replacement of the secondary bromide or (10) L. N. Owen and P. A. Robins, J . Chem. S O ~ .326 ,
.----,-
(5) S. F. Birch, T. V. Cullum, and R. A. Dean. Paper presented before the Division of Petroleum Chemistry, 130th Meeting of the American Chemical Society, Atlantic City, New Jersey, September, 1956. (6) 5. F. Birch, R. A. Dean, and E. V. Whitehead, J . Org. Chem., 19, 1449 (1954). (7) Di-ptoluenesulfonate. (8) . . F. S. Fawcett. Chem. Revs.. , 47., 219 (1950).
.
,
(1949).
(11) L. N. Owen and P. A. Robins, J . Chem. SOC.,320 (1949). (12) M. F. Clarke and L. N. Owen, J . Chem. SOC.,2103 (1950). ’ (13) B. Rothstein, Ann. Chem., 14, 461 (1930). (14) H. Pines, A. Rudin, and V. N. Ipatieff, J . Am. Chem. Soc., 74.4063 (1952). (15) H. Phillips, Chem. SOC., 2552 (1925).
>.
1592
BIRCH, DEAN, HUNTER, AND WHITEHEAD
route and for these reasons, 2-thiabicyclo[3.2.1]octane (IX) was prepared by this route, as shown in the reaction scheme (V-X). The alcohol VI was obtained in good yield when the ester V (prepared by the method of Noller and AdamsI6) was reduced with lithium aluminum hydride. v. Uraun et al." reported that on hydrobromination of VI they obtained 1-bromo-3-(2bromoethy1)cyclopentane (VII) and not the corresponding 2-(2-bromoethyl) compound (VIII). However, even if isomerization did not occur under the conditions employed by these workers, it seems unlikely that VI1 was the only addition product, and certainly their identification of their dibromide as this compound does not appear conclusive. We anticipated that the corresponding 1- and 242bromoethyl) compounds would accompany 1%romo-3-(2-bromoethyl)cyclopentanein our preparation, and the decomposition, accompanied by evolution of hydrogen bromide, which occurred on distillation of our hydrobromination product, and the occurrence of 2-thiabicyclo [3.3.01octane in addition to the required 2-thiabicyclo[3.2.1]octane in the product obtained on treatment of the distilled dibromide with sodium sulfide, confirmed our expectations. Fractional distillation of the mixture of bicyclic sulfidos (34% yield) separated from the polymeric portion of the product, followed by fractional crystallization of the mercuric chloride complexes of the liquid and solid fractions so obtained, gave a quantity of the 2-thiabicyclo[3.2.1]octane complex, and the sulfide regenerated from this was purified by crystallization and sublimation. The sulfide regenemted from the combined residual complexes was shown hy spectroscopic examination to contain ds-2-thiabicyclo[3.3.0]octane; no trrm,s-2-thiabicyclo[3.3.O~octanelB could be detected, however, and it seems that the amount of the cis-isomer t f VTII in the hydrotirornination product was very small. The diols required for the synthesis of 6-thiabicyclo[3.2.l?octane (XIII) and 2-thiabicyclo[2.2.2]octane (XVII), uiz, the 3- and 4-hydroxycyclohexanemetbanols IXX and XIV), are readily obtainable, arid XI11 and XVTI were therefore synthesized b y the ditosylate route as indicated schematically KI-XVII). These diols were prepared, in average yields of 69 and 63a/, respectively, as describtd hy Owen et U Z . ~ , ~ ONeither trans-3-hy(3roxycyelohcxanemethanol nor its ditosylate can readily hc separated from their respective admixtures that,a mixwith the corrwpondingcis-comp~unds,~so turc of cis- and trans-ditosylates, obtained in almost theoretical yield hy the usual method,lghad to be (16) C>. R. Noller and R. Adams, J. Am. Chem. Soc., 48, 2444 (1926). (17) J. v. Rraun, E. Kamp, and J. Xopp, Ber., 70, 1750 (1937). (18) S. F. Birch, R. A. Dean, and E. V. Whitehead, Chemistry & Industry (London), 409 (1956) and a future
publication in this series.
VOL.
22
C-C-CH2OH cis
+ trans
cis
pTsCl
Cryat.
C-C-OTS
i
+ trans
1
(CHs)z C=O/EtzO
CC-OH
l
C-C-CH~OTS cis
+ trans
4
NarS
6--cc
(cis)
trans pTsCl
C-C-OTs
trans NatS
c--cs
ic1
(XVII)
t - L C (Cis)
used for the preparation of XIII. The trans- isomer of 4-hydroxycyclohexanemethanol is easily separated from its admixture with the cis- isomer'o however, so that the pure trans-ditosylate (XVI) was used for the preparation of XVII. Only a poor yield of XVI was obtained in a preliminary preparation carried out in the usual way,19 and, following a suggestion by Haggis and Owen,2othe majority of this ditosylate was prepared by reverse addition of the reactants. The yields of XI11 and XVII obtained on reaction of these ditosylates with sodium sulfide were very low (ca. 5%), the majority of the product being di- or poly-meric material. Consequently, since it appeared likely that the dibromide corresponding to the structure of the original diol would be the main hydrobromination product, it was considered worthwhile to investigate whether, despite the disadvantages already mentioned, the dibromide route might not prove to be on the whole more satisfactory than the ditosylate route. The yields of bicyclic sulfide were certainly increased to 15 and 12% based on the dibromides obtained from XI and XIV respectively, but surprisingly, this sulfide consisted in both instances almost entirely (19) E. c. Homing, Org. Syntheses, Coll. Vol. 3, 366 (1955). (20) G. A. Haggis and L. N. Owen, J. Chem. SOC., 389 (1953).
DECEMBER
1957
1593
SULFUR COMPOUNDS RELATED TO PETROLEUM. VI1
TABLE I MELTINGPOINTS AND ANALYSESOF THE DERIVATIVES OF
THE
SULFIDES
Analyses Formula of derivative
Compound 2-Thiabicyclo [3.2.1 ]octane 6-Thiabicyclo [3.2.1 ]octane 8-Thiabicyclo [3.2.1]octane 2-Thiabicyclo [2.2.2]octane
" With decomposition.
M.P., "C., (corrected)
193-194' C7HiaClzHgS 257" C7H'20zS 210-211"*b C8Hi.JS CdhJX"S 157-157.5" C~H~ZOZS 236-237 CsHljIS 162.5-163.5'9' CTH~ZC~ZH~S 226= CIHIZO~S 281.5-282 GHiJS 242.5-243"~~ C~HiaCLHgS 185-1 87" C~H~ZOZS ca. 310 CsHidS 225-226"~~
C
Calcd. H
S
C
Found H
S
21.0 52.5 35.6 21.0 52.5 35.6 21.0 52.5 35.6 21.0 52.5 35.6
3.0 7.6 5.6 3.0 7.6 5.6 3.0 7.6 5.6 3.0 7.6 5.6
8.0 20.0 11.9 8.0 20.0 11.9 8.0 20.0 11.9 8.0 20.0 11.9
21.3 52.3 35.7 21.1 52.8 35.7 20.9 52.3 35.6 21.3 52.6 34.9
3.0 7.7 5.8 3.3 7.8 5.6 3.0 7.8 5.8 3.2 7.7 5.3
8.1 20.1 12.0 8.2 20.4 12.4 8.0 20.2 12.0 7.9 20.1 12.3
Sealed tube plunged into bath a t ca. 10" below m.p.
of XIII, and XVII was not present in detectable quantities in either product. This would seem to show that while with X I the hydrobromination complexes and reaction products do not tend to rearrange in such a way that truns-l-bromo-4(bromomethy1)cyclohexane is formed, with XIV extensive rearrangements occur such that truns-lbromo-3-(bromomethyl)cyclohexane is a major end product. Thus, while a quantity of XI11 was obtained by this dibromide route, further amounts of XVII had to be prepared via the ditosylate. Specimens of XI11 and XVII were purified by crystallization of their mercuric chloride complexes, followed by crystallization and sublimation of the regenerated sulfides. All four sulfides are, like 3-thiabicyclo[3.2.1]octane, wax-like solids with strong camphoraceous odors. 6- and 8-Thiabicyclo[3.2.1]octane melt a t about the same temperature as 3-thiabicyclo[3.2.1]octane (176175") and the melting point of 2thiabicyclo [3.2.1] octane is only about 10" lower, while the melting point of 2-thiabicyclo[2.2.2]octane is as high as 210-212". Melting points of mixtures of the sulfides are indefinite, but are intermediate between those of the components of the mixtures, Le. there is no true depression of the melting point. All five sulfides of the group have boiling pointsz1 within two or three degrees of 195" a t atmospheric pressure. (The boiling points of 3-thiabicyclo[3.2.l]octane and 2-thiabicyclo[2.2.2]octane21have not actually been determined, but their values can be deduced from data on the occurrence of the sulfides in fractions obtained on distillation of sulfide concentrates isolated from crude keroseneb.) Derivatives of the four sulfides described in this
paper were prepared by methods described previously6,22and their melting points are given in (21) Boiling point defined as temperature a t which vapor pressure is 760 mm. 2-Thiabicyclo [2.2.2]octane sublimes when heated a t atmospheric pressure and has no normal boiling point. (22) E. V. Whitehead, R. A. Dean, and F. A. Fidler, J . Am. Chem. SOC.73,3632 (1951).
Table I ; the high melting points of the sulfones, especially that of 2-thiabicyclo[2.2.2]octane, are. noteworthy. The infrared spectraz3of these sulfides and of 3-thiabicyclo[3.2. lloctane have been obtained in the range 2-15 p using a Grubb Parsons double beam spectrometer. EXPERIMENTAL
All melting points are corrected. Microanalyses are by Dr. Ing. A. Schoeller of Kronach/Oberfranken, Bambergerstrasse 20, Germany. ' 8-Th~icycZo[8.%.l]S-octae-8,8dioxide (11) was obtained in 90-95y0 yield when 1,3-cycbheptadiene (I) (25 g.) (b.p. 120-121 "/770 mm., n z 1.4960), phenyl-&naphthylamine (0.2 g.), and sulfur dioxide (86 g.) were reacteda a t room temperature for 40 hr. "he product, worked up in the usual way2 and crystallized from ethanol, melted a t 160.5-162'. Anal. Calcd. for C ~ H I ~ C, O ~53.1; : H, 6.4; S, 20.3. Found: C, 53.2; H, 6.5; S, 20.3. 8-Thiabieyclo[8.%.1]oclans8,8diode (111) was obtained in theoretical yield by hydrogenating2 11. The product was not distilled, but was crystallized from ether or benzene/ cyclohexane. It melted a t 281.5-282'. 8 - T h i a b ~ c ~ [ S . d . l ] o c ~(IV). n e Reductione of the sulfone (111) (1 mole) with lithium aluminum hydride (2 moles) gave 66% of the sulfide (IV) and 30% of unreduced sulfone (111). Crystallized to constant melting poiat from n-pentane and sublimed a t 0.4 mm., S-thiabicyclo[3.2.l]octane melted (sealed tube) a t 176.5-178.5", b.p. 194.5'/769 mm. Anal. Calcd. for C,HI2S: C, 65.6; H, 9.4; S, 25.0. Found: C, 65.6; H, 9.5; S, 25.0. d-Cyclopenhe-l-acettk acid was obtained in 65% yield by the method of Noller and Adams.lB The diester was not isolated but was hydrolyzed and decarboxylated to give the required acid (b.p. 90°/2 mm.-90'1.2 mm., n v 1.4682). The ester (V) obtained in 82% yield by continuous esterification of the acid, boiled at 95"/29 mm.-96"/27 mm., n'," 1.4480. .?-Cyclopatene-l-ethanol (VI) was prepared by reducing the ester (V) (258 9.) with lithium aluminum hydride (38 g.), yield 87%, b.p. 98"/30 mm., nv 1.4722. (Hydrogenation of a portion of VI with a platinum black catalyst gave cyclopentanethanol in 87% yield, b.p. 96"/25 mm., n y 1.4577). Bromo-(d-bronwethyl)eyclopentanes (VII, VIII). Anhydrous hydrogen bromide was passed into a mixture of the alcohol (VI) (108 g.) and water (1 g.), a t a maximum tem(23) These spectra have been submitted to the A.P.I. Research Project 44 for inclusion in their catalog of spectral data.
1594
BIRCH, DEAN, HUNTER, AND WHITEHEAD
VOL.
22
perature of 70' until no further increase in weight occurred. After working up the product in the usual w a p it was distilled under reduced pressure when partial decomposition accompanied by evolution of hydrogen bromide took place. The following fractions were obtained: 1. b.p.
